sizeof module

Simple function for determining the memory usage of common Python values and objects.

sizeof.sizeof.arch() → int

Determine the architecture.

>>> arch() in [32, 64]
True

sizeof.sizeof.sizeof(obj: Any, deep: bool = False) → int

Estimate the memory consumption of a Python object (either the root object itself exclusively or, in the case of a deep traversal, the entire tree of objects reachable from it).

>>> sys.getsizeof([]) == sizeof([])
True
>>> sys.getsizeof(123) == sizeof(123)
True
>>> sys.getsizeof(123.0123) == sizeof(123.0123)
True
>>> sys.getsizeof(complex(2, -3)) == sizeof(complex(2, -3))
True
>>> sys.getsizeof(True) == sizeof(True)
True
>>> sys.getsizeof(None) == sizeof(None)
True
>>> sizeof('ab') == sizeof('a') + 1
True
>>> sizeof(bytes([1, 2])) + 1 == sizeof(bytes([1, 2, 3]))
True
>>> sizeof(bytearray([1, 2])) <= sizeof(bytearray([1, 2, 3]))
True
>>> mv0 = memoryview(bytes([1, 2]))
>>> mv1 = memoryview(bytes([1, 2, 3]))
>>> sizeof(mv0) == sizeof(mv1)
True
>>> xs = [1, 2, 3]
>>> ys = {'a':1, 'b':2, 'c':3}
>>> zs = {frozenset([1, 2, 3]): [1, 2, 3], frozenset(['a']): 'bc'}
>>> sys.getsizeof(xs) == sizeof(xs)
True
>>> sys.getsizeof(range(0,10)) == sizeof(range(0,10))
True

When the deep argument is True, the total amount of memory consumed by the object and all of its descendants by reference is calculated (with deduplication).

>>> sys.getsizeof(xs) < sizeof(xs, deep=True)
True
>>> sizeof(set(xs), deep=True) > sys.getsizeof(set(xs))
True
>>> sizeof({123}) == sys.getsizeof({123})
True
>>> sizeof({123}, deep=True) > sys.getsizeof({123})
True
>>> sizeof([xs], deep=True) > sys.getsizeof([xs])
True
>>> sizeof([xs, ys, zs], deep=True) > 2 * sizeof([xs, xs, xs], deep=True)
True
>>> sizeof((xs, ys, zs), deep=True) > 2 * sizeof((xs, xs, xs), deep=True)
True
>>> sizeof([xs, xs], deep=True) < sizeof([xs, [1, 2, 3]], deep=True)
True
>>> sizeof([xs], deep=True, _exclude=set([id(xs)])) == sizeof([xs])
True
>>> sizeof(range(0,10), deep=True) > sizeof(range(0,10))
True